Soil denitrification fluxes and oxygen dynamics in three contrasting northeastern North American forests

Background/Question/Methods

Denitrification, the microbially mediated conversion of nitrate to nitrogen (N) gases (NO, N₂O, and N₂), represents a major pathway for ecosystem N loss under anoxic conditions, but its significance in most ecosystems, especially those dominated by oxic conditions is unclear. We measured denitrification in soils with varying oxygen (O₂) regimes and N availability in three northeastern forests: 1) Hubbard Brook Experimental Forest (HBEF), New Hampshire, USA; 2) Turkey Lakes Watershed (TLW), Ontario, Canada; and 3) Bear Brook Watershed (BBWM), Maine, USA. We installed dataloggers with soil oxygen, moisture and temperature sensors to monitor these variables in the top 5 cm at hourly intervals during the growing season (June to September 2011). In each forest, we exploited the range of environmental conditions available: in HBEF, we monitored six plots along an elevation and moisture gradient; in TLW, we monitored six plots along two hillslope to wetland transects; and in BBWM, we monitored 4 plots, in hardwood and softwood plots in N-fertilized and control watersheds. We collected soil samples (0-10 cm; varying proportions of organic and mineral soil) in June and September/October in all plots and measured denitrification gas fluxes under varying He-O₂atmospheres. We also measured soil characteristics and microbial activity. 

Results/Conclusions

We found substantial denitrification gas fluxes at 10% and 20% O₂ in lab incubations. We expected to find significant variations in soil O₂ content in response to rain events, but found that only wetland sites in TLW had O₂ concentrations significantly below 20%. In the high-low elevation contrast at HBEF, we found higher soil moisture, respiration, and denitrification rates in the high elevation sites. In the hillslope-wetland transects at TLW, we measured higher respiration, %O₂, and N₂O fluxes (up to 1.2 µg N/g soil /day), but lower denitrification rates, in the toe slopes compared to wetland sites. In the hardwood-softwood and fertilization contrasts in BBWM, we found no clear patterns: across plots, there was no difference in %O₂, but hardwood plots had lower respiration and denitrification rates, regardless of N fertilization, while fertilized softwood plots had high N₂O fluxes (up to 0.3 µg N/g soil /day). Denitrification fluxes were generally highest at 10% O₂ in all forest sites, but we did not find consistent relationships between denitrification fluxes and indices of N and C availability among forests.  Denitrification appears to be a significant sink for N in these forests, but the factors controlling this process remain unclear.